Update Nightly (June 2018)

[bradjc]

Pull Request Overview

This pull request updates rust to a recent version.

That also means a new rustfmt, which is annoying. The commits are more easily viewed independently.

Testing Strategy

This pull request was tested by running hail on hail.

TODO or Help Wanted

n/a

Documentation Updated

• Kernel: Updated the relevant files in /docs, or no updates are required.
• Userland: Added/updated the application README, if needed.

Formatting

• Ran make formatall.

[phil-levis]

Should we get into the habit of bumping version after each release?

[bradjc]

Should we get into the habit of bumping version after each release?

We pretty much are already doing that by updating the nightly every two months. And I think it makes sense to do it after a release so there is time for people to use it and test it.

This is blocked on tonight's nightly which should include binutils.

[bradjc]

Ok! This is ready to go now.

[bradjc]

And travis only takes 5m versus 19m for current pull requests.

[bradjc]

Anyone not ok with merging this? I've been waiting for a nightly to land that includes llvm-tools, but otherwise this hasn't changed much from 12 days ago. Included in this PR:

• Use nightly-2018-06-26.
• Remove the install rustfmt script and instead use rustup component add rustfmt-preview. This solves two problems: 1) we don't have to compile rustfmt any more :))) and 2) rustfmt is tied to the version of rust so no more guessing about which rustfmt version will work with which nightly. It also has an additional benefit: travis is much faster.
• Actually run rustfmt, which makes a ton of minor changes. I guess that is the price of progress and convenience. There is a milestone and tracking issue for rustfmt 1.0: https://github.com/rust-lang-nursery/rustfmt/milestone/2. As far as I can tell 1.0 would guarantee that any format specified in the RFC that is eventually merged wouldn't change, but that unspecified format could still change in the future. I also get the sense that deciding that that RFC is done and ready to merge is the real blocking issue.